To date, the discrete short time Fourier transform (DSTFT) and similar analysis techniques have generally been used to analyze only individual notes. This paper discusses problems (and their solution) in using the DSTFT for analyzing transitions between performed notes. Recordings of transitions performed on the trumpet, clarinet, and violin were analyzed with the DSTFT. Based on the responses from 10 musically sophisticated subjects, the DSTFT was shown to be adequate for modeling transitions. To create line-segment approximation of the original DSTFT data, various semi-automatic methods were developed or adapted from the literature on pattern recognition and approximation theory. In a second experiment with the same subjects, line-segment approximations were shown to model musical transitions adequately.
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.